Wolfgang Denk wrote:
In message 46B9563B.1010807@compulab.co.il you wrote:
Could this ever be other than the first Ethernet controller on a board? If so, 'ethaddr' won't cut it.
If a board has more than one DM9000, the driver in its current state does not support it anyway.
Assume the board has an XXX Ethernet controller as eth0, and a DM9000 as eth1 ?
The only solution I can think of in this case is adding some #define CONFIG_DM9000_ETHER_INDEX in the board config and approptiate '#if' clauses in the driver. If it's Ok with you I'll do it and resend the patch. Still, the configuration you mention seems very theoretical for me and I can hardly imagine someone designing a board with such configuration.
Please choose a better name for this function. While multicast addresses are poor choices for source, they're perfectly fine for destination.
What about 'is_valid_ether_src_addr'?
That's a very long name which is difficult to read, difficult to type, and makes for very long source code lines...
And I will probably have to look up the definition of this function each time I read the code. You use this function exactly once, so I suggest to NOT define it at all and write down what it does:
if (!is_multicast_ether_addr(addr) && !is_zero_ether_addr(addr)) ...
This is IMHO much better to read and to understand.
I agree and I'll fix it.
Best regards,
Wolfgang Denk